An effective Hilbert-Huang transform-based approach for dynamic eccentricity fault diagnosis in double-rotor double-sided stator structure axial flux permanent magnet generator under various load and speed conditions

Eccentricity fault in double-sided axial flux permanent magnet generator is very difficult to be detected as the generated variations terminal electrical parameters are weak and chaotic, especially at initial stages of occurrence. In addition, one most important problems any diagnosis approach investigation load speed variation on proposed indices. To overcome aforementioned difficulty problems, this paper adopts a novelty detection algorithm based Hilbert-Huang transform (HHT) which time-frequency signal analysis empirical mode decomposition Hilbert transform. It well suited for reliable since it unaffected by transient conditions make process incur into false alarms. The HHT-based methods has been demonstrated recent years rotor bearing faults induction machine also stator identification PM synchronous machines with radial structure. This study explores possibility applying technique dynamic eccentricity double-rotor structure under variable conditions. relies two steps: estimating intrinsic functions (IMFs) (EMD) computing instantaneous amplitude (IA) frequency (IF) IMFs using more significant determined means spectrum, applied accurate diagnosis. severity can evaluated energy value. theoretical basis method presented. effectiveness verified series simulation experimental tests different results show that presented robust against variations.